Charles Hanley

The "Difficulty" of a Personality Inventory Item

ployed with tests. Fortunately, it is well-known in psychophysics (Woodworth, 1938) that the more difficult a judgment is to make, the longer it takes. (When two weights are nearly equal, for example, deciding which is heavier is slower work than when the weights differ markedly.) If judging the personal applicability of inventory statements is similar to judging physical dimensions of stimuli, difficult personality items will have longer response latencies. Similarly, just as the more difficult psychophysical judgments are accompanied by lowered confidence by the judge in his decision, so difficult personality items should be associated with lowered confidence by the individual in his answers. The present study involves a first-approximation to the measure-

Size constancy in extended haptic space

When a stick is used to extend the boundaries of the world perceived by touch, haptic parallels to visual angle and size constancy appear. In a pilot study, Ss using a stick to feel an out-of-reach gap were more likely to underestimate its size than were Ss feeling a nearby gap with the index finger, but size constancy was relatively high with the stick. The main study, comparing judgments of gaps at intermediate and far positions, confirmed these findings. Errors were greater for the more distant gap, but constancy was high in both positions. Correlations between far and intermediate judgments indicated that Ss were consistent in the type of error made. Similar findings emerged from a partial replication with blind and partially sighted Ss.

A note on the up-and-down method

Objections to the use of the original version of the up-and-down method in psychology, as exaggerating individual differences and yielding discrepant results, overlook statistical, psychometric, and logical factors. When the method is applied incorrectly, the standard error of SD is largeand apparent differences in estimates of SD are not statistically significant. Psychometrically, a singletrial procedure should give less reliable, hence more variable. PSEs than one using multiple trials. But this effect increases standard error of the mean. thus significant mean differences found with the method are not attributable to greater variability in PSEs. Explanations of discrepancies in results from up-and-down and traditional methods in the study of perceptual development fail to account for numerous similarities. Finally, a novel method ought to produce some novelty, if it is to make a contribution to psychology,

Relationship among PGR indices

Abstract Psychogalvanic reflex (PGR) has been widely employed in psychological research as an index of emotionally, autonomic responsivity, sub-clinical response etc., and has been an integral part of research design investigating phenomena as disparate as perceptual vigilance and psychosomatic illness. Little attention is paid, however, to the identity of PGR as a tool; studies tend to regard, and therefore employ, PGR as a unitary measure devoid of facets which in themselves relate to planning of a research design. The present study employed seven separate measures on 64 normal subjects with mean age of 20·5 yr, four measures relating to different types of stress stimulation and three different measures of the resting phase. While some measures correlated highly with each other, others failed to correlate or showed high negative correlations. Three separate PGR factors were obtained. Resulsts emphasize a need for (1) further investigation of PGR as a research tool and (2) judicious consideration of type of PGR measurement employed with respect to given research expectations.